cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Steelers_In_DC
‎07-12-2025
  • 891 Posts
  • 36 Likes Given
  • 19 Solutions
  • 213 Likes Received
About
I have worked on a research team, as a data/risk analyst, also as a Marketing Analytics Manager, as a consultant and now in the healthcare industry.

Re: Not enough memory for proc transpose

by Barite | Level 11 in SAS Procedures
‎04-20-2016 10:23 AM
‎04-20-2016 10:23 AM
Here's what I get when I limit the population in the first step.  This is what I want:   run_date reporting_month Rule Fail_Total 04/20/2016 01/31/2016 1_All Failures   04/20/2016 01/31/2016 2_All Failures 0 04/20/2016 01/31/2016 2.1_All Failures 0 04/20/2016 01/31/2016 2.2_All Failures 0 04/20/2016 01/31/2016 2.4_All Failures 0 04/20/2016 01/31/2016 3_All Failures 11 04/20/2016 01/31/2016 3.1_All Failures 0 04/20/2016 01/31/2016 3.2_All Failures 0 04/20/2016 01/31/2016 3.3_All Failures 11 04/20/2016 01/31/2016 3.5_All Failures 0 04/20/2016 01/31/2016 4.1_All Failures   04/20/2016 01/31/2016 5.1_All Failures   04/20/2016 01/31/2016 5.2_All Failures 0 04/20/2016 01/31/2016 6.2_All Failures 145 04/20/2016 01/31/2016 6.3_All Failures 0 04/20/2016 01/31/2016 7.2_All Failures 0 04/20/2016 01/31/2016 8.1_All Failures 23 04/20/2016 01/31/2016 8.2_All Failures   04/20/2016 01/31/2016 8.4_All Failures 1595 04/20/2016 01/31/2016 9_All Failures 0 04/20/2016 01/31/2016 9.1_All Failures   04/20/2016 01/31/2016 10.2_All Failures 0 04/20/2016 01/31/2016 11_All Failures 98919 04/20/2016 01/31/2016 11.1_All Failures 0 04/20/2016 01/31/2016 11.3_All Failures 51 04/20/2016 01/31/2016 12_All Failures 98765 ... View more

Re: Not enough memory for proc transpose

by Barite | Level 11 in SAS Procedures
‎04-20-2016 10:22 AM
‎04-20-2016 10:22 AM
Nope, that's it.  Sounds remedial when I say it but I'm not sure what to do. ... View more

Re: Not enough memory for proc transpose

by Barite | Level 11 in SAS Procedures
‎04-20-2016 10:11 AM
‎04-20-2016 10:11 AM
I have several datasets that have edits that need to be applied for governmental checks.  I'm working on the largest one now which is 1st lien mortgage data.  I wrote out 350+ rules like this:   '_1_All Failures'n =1; '_2_All Failures'n = missing(loan_closing_date); '_2.1_All Failures'n = loan_closing_date > reporting_month; if not missing(loan_closing_date) then do; '_2.2_All Failures'n = loan_closing_date < Forty_years_ago;end;   I chose the name of the rules to coincide with a name provided by the government so I can join and get additional information downstream.  This produces a dataset with 558 variables by 987,144 rows. ... View more

Not enough memory for proc transpose

by Barite | Level 11 in SAS Procedures
‎04-20-2016 09:31 AM
‎04-20-2016 09:31 AM
Good Morning All,   I have a solution that I've used in the past with smaller datasets, the current dataset I am working with has 350 'rules' and about 1 million records, so the transpose step isn't working due to the wide dataset that is produced.  I tried to split the original dataset up into 3rds but the 300,000 columns are still too much.  Below is a short example of what I am doing, any suggestions or different solutions are greatly appreciated.   data have; infile cards dsd; informat run_date reporting_month mmddyy10.; format run_date reporting_month mmddyy10.; input run_date reporting_month    _rule1 _rule2 _rule3; cards; 04/20/2016,03/31/2016,.,1,0 04/20/2016,03/31/2016,1,1,1 04/20/2016,03/31/2016,1,1,1 04/20/2016,03/31/2016,1,1,0 04/20/2016,03/31/2016,.,0,0 ;run; proc transpose data=have out=tran_have;by Run_Date reporting_month;var _:; data want(keep=Rule Fail_Total Run_Date reporting_month); set tran_have; Rule = substr(_NAME_,2,5); Fail_Total = sum(of col:); run; ... View more

Re: check field for valid state name

by Barite | Level 11 in SAS Programming
‎04-05-2016 08:45 AM
‎04-05-2016 08:45 AM
Other than using a macro variable so I don't have a long in() clause I have no idea.  Looking for anything to be more efficient.  There's a large amount of records that I need to check. ... View more

check field for valid state name

by Barite | Level 11 in SAS Programming
‎04-01-2016 03:24 PM
‎04-01-2016 03:24 PM
I am checking a field for a valid postal code.    'Property State' IN ( "AL", "AK", "AZ", "AR", "CA", etc...   Is setting this up in a macro faster than having it in an if/then statement?  I'm checking a file with a couple million rows and want to make sure I'm efficient as possible.   Thanks in advance, have a great weekend. ... View more

Re: verify number and length

by Barite | Level 11 in SAS Procedures
‎04-01-2016 08:19 AM
‎04-01-2016 08:19 AM
Here's is what I have that I like better than my original code:   data data; informat one 16.; format one 16.; input one; cards; 1 1.21 1000 1234567891123456 12345678911234567 ;run; data qc; set data; if not missing(scan(one,2,'.')) then check = 1; else check = 0; run;   the etl will only load a number, the long number goes to 1.E so scan will pick it up, the decimal will get picked up using the scan also.  I don't like that the 1.21 doesn't show in the output but it flips the error so I think it's acceptable. ... View more

Re: verify number and length

by Barite | Level 11 in SAS Procedures
‎04-01-2016 08:06 AM
‎04-01-2016 08:06 AM
It's not a string, I have to verify it's a number with a length of 16.0.  The number part will be a verified at load time, I'm using the length function to verify the length and the scan function to make sure there are no numbers after a decimal point. ... View more

verify number and length

by Barite | Level 11 in SAS Procedures
‎04-01-2016 06:57 AM
‎04-01-2016 06:57 AM
I have a numer of fields that I need to verify are a number with a length smaller than 17 and no decimal places.  Through my ETL I'm verifying it's a number, to verify length I'm using strip(length()), and to verify there are no decimal places I'm using scan().  Is there a more efficient way of doing this?   Thank You, ... View more

Re: proc gchart warnings

by Barite | Level 11 in Graphics Programming
‎03-29-2016 12:44 PM
‎03-29-2016 12:44 PM
I'm not sure what you mean by 'dirty'.  Do you mean because it's monthly, or because of the way the pass/fail is set up.  I'm not sure how to clean it up.  ... View more

proc gchart warnings

by Barite | Level 11 in Graphics Programming
‎03-28-2016 10:50 AM
‎03-28-2016 10:50 AM
Below is a sample dataset and code that I am running.  If you run this you will see the legend says Fail, 0,1.  I would like it to show the total of passes and fails.  So it would have pass 9, fail 1.  Any help will be appreciated. Also, I'm not experienced using graphs, so any other suggestions or pointers are welcome.   edit*  Also, I am getting the following warnings when I go to my full datasset:   WARNING: The intervals on the axis labeled "REPORTING_MONTH" are not evenly spaced. WARNING: Some INSIDE= labels were not drawn due to space limitations.   Thank You,   data data; infile cards dsd; input reporting_month$ rule$ fail_total column$ description$ pass fail; cards; 20151031,0001,0,BHC_NAME,No Nulls,1,0 20151130,0001,0,BHC_NAME,No Nulls,1,0 20151231,0001,0,BHC_NAME,No Nulls,1,0 20151031,0002,0,ID_RSSD,No Nulls,1,0 20151130,0002,0,ID_RSSD,No Nulls,1,0 20151231,0002,0,ID_RSSD,No Nulls,1,0 20151031,0003,0,PORTFOLIO_ID,ContentCheck=USOthCons,1,0 20151130,0003,0,PORTFOLIO_ID,ContentCheck=USOthCons,1,0 20151231,0003,0,PORTFOLIO_ID,ContentCheck=USOthCons,1,0 20151031,0004,0,REPORTING_MONTH,ContentCheck=QTR,1,0 20151130,0004,0,REPORTING_MONTH,ContentCheck=QTR,1,0 20151231,0004,0,REPORTING_MONTH,ContentCheck=QTR,1,0 20151031,0005,0,SEGMENT_ID,No Nulls,1,0 20151130,0005,0,SEGMENT_ID,No Nulls,1,0 20151231,0005,0,SEGMENT_ID,No Nulls,1,0 20151031,0006,0,PRODUCT_TYPE,ContentCheck=01-05,1,0 20151130,0006,0,PRODUCT_TYPE,ContentCheck=01-05,1,0 20151231,0006,0,PRODUCT_TYPE,ContentCheck=01-05,1,0 20151031,0007,0,PRODUCT_TYPE,Comapre01=Secured/Revolving,1,0 20151130,0007,0,PRODUCT_TYPE,Comapre01=Secured/Revolving,1,0 20151231,0007,0,PRODUCT_TYPE,Comapre01=Secured/Revolving,1,0 20151031,0008,0,PRODUCT_TYPE,Compare02=Secured/Installment,1,0 20151130,0008,0,PRODUCT_TYPE,Compare02=Secured/Installment,1,0 20151231,0008,0,PRODUCT_TYPE,Compare02=Secured/Installment,1,0 20151031,0009,0,PRODUCT_TYPE,Compare03=Unsecured/Revolving,1,0 20151130,0009,0,PRODUCT_TYPE,Compare03=Unsecured/Revolving,1,0 20151231,0009,0,PRODUCT_TYPE,Compare03=Unsecured/Revolving,1,0 20151031,0010,1,PRODUCT_TYPE,Compare04=Unsecured/Installment,0,1 20151130,0010,1,PRODUCT_TYPE,Compare04=Unsecured/Installment,0,1 20151231,0010,1,PRODUCT_TYPE,Compare04=Unsecured/Installment,0,1 ;run; proc gchart data=data;                                                                                                                      vbar reporting_month / discrete subgroup=fail                                                                                                         /*group=pass*/ /*g100 nozero */                  /*sum=fail*/ /*type=sum  */                                                                                                           inside=freq width=10;*                                                                                                                gaxis=axis1 raxis=axis2;*                                                                                                                legend=legend1; run;                                                                                                                                     quit;  ... View more

Re: Proc Print - how to format output

by Barite | Level 11 in ODS and Base Reporting
‎03-21-2016 04:31 PM
‎03-21-2016 04:31 PM
I think this is what you are looking for:   data have; input one$ two$; cards; one two ; run; data want; set have; one = catx(' ',one,two); run; ... View more

if then efficiency / options

by Barite | Level 11 in SAS Programming
‎03-11-2016 06:55 PM
‎03-11-2016 06:55 PM
I have the code below, it works fine but with a larger dataset it will be innefficient.  If anyone can help get the same results, I want to do something if else/then but am not sure how to do it with the multiple filters.  The segment_id and product_type will change each line, but I only care about the one per line.  The selection should account for the entire population.   _0005 = missing(segment_id); if substr(segment_id,1,2) in ('01','02','03','04','05') then do; _0006 = 0;else _0006 = 1; if substr(segment_id,1,2) = '01' and product_type = 'Secured-Revolving' then _0007 = 1;else _0007 = 1; if substr(segment_id,1,2) = '02' and product_type = 'Secured-Installment' then _0008 = 0;else _0008 = 1; if substr(segment_id,1,2) = '03' and product_type = 'Unsecured-Revolving' then _0009 = 0;else _0009 = 1; if substr(segment_id,1,2) = '04' and product_type = 'Unsecured-Installment' then _0010 = 0;else _0010 = 1; if substr(segment_id,1,2) = '05' and product_type = 'Overdraft' then _0011 = 0;else _0011 = 1; ... View more

Proc Format or if/then, or any other ideas?

by Barite | Level 11 in SAS Programming
‎03-11-2016 09:19 AM
‎03-11-2016 09:19 AM
I have a file with 675 records, for this file I have 64 rules or conditions that need to be applied, such as ID is not null, Account_Number is a length of 16, balance isn't longer than 16.6.   I will get a fail rate from month to month for each of the 64 rules, so I'm thinking of writing 64 if/then statements, which will generate 64 new binary columns, or 64 formats that will generate 64 new binary columns.  This doesn't sound appealing.  If I get through this process and am able to automate it I will then move on to much larger files that have much larger sets conditions checked.  The rule and the fail rate is all I am concerned about to start.  Towards the end of the project I will need to drill down and find out at an account level why something failed.    This is currently in an application called Trillium, which works fine but is vary manual, my task is to automate the process.    Any and all thoughts and suggestions are welcome.  Thank You, ... View more

Re: Merge/join

by Barite | Level 11 in SAS Programming
‎03-03-2016 12:31 PM
‎03-03-2016 12:31 PM
I don't know what ALL form ids or questions means.  I need a list of how many there are and what they are.  ... View more
Contact Me
Online Status
Offline
Date Last Visited
‎07-12-2025 12:12 PM
Likes given to
Likes from
Quick links